In the first section we describe mutagenesis of the proposed substrate binding domains of the viral protease NS3. Last year we described the construction of an infectious full-length cDNA clone of DEN2, and further characterization of this clone is described in the middle section. Finally, we describe the introduction of selected mutations in NS2B into the DEN2 genome. Mutagenesis of NS3. Two of the four domains in NS3 with sequence homology to trypsin-like serine proteases ("homology boxes" 3 and 4) contain residues predicted to be important for substrate binding (aa 1604, 1605, 1625, 1627 and 1628). We have begun a limited mutagenesis of some of the residues in these two boxes in order to characterize their requirement for protease activity and to identify mutants with partial cleavage defects. To date, 34 mutations have been analyzed. Most mutations in box 4, including those in aa 1625, 1627, and 1628, abolish protease activity. This is consistent with the hypothesis that box 4 forms part of the substrate-binding pocket. In contrast, many of the mutations in box 3, including those in aa 1604 and 1605, retain significant protease activity, even wild-type activity in a few cases. Surprisingly, asp 1604, predicted to contact a positively charged residue at the cleavage site, can be changed to lys or arg with significant residual activity. Analogous mutations in trypsin abolish cleavage. Other mutations in box 3, clustered around the catalytic serine, abolish protease activity. This confirms that some resisues in box 3 are important for protease function, but is inconsistent with the proposed role for aa 1604 and 1605 in substrate binding. Control mutations outside of the homology boxes were well-tolerated. Our plan is to introduce some of the mutations in NS3 which permit significant residual protease activity back into the dengue genome, in an effort to confirm the observed effect on protease function and possibly to attenuate the virus. Characterization of the DEN2 infectious clone. By plaquing directly after RNA transfection we find that the specific infectivity of transcripts is 10-100 fold less than virion RNA. The DNA sequence of our infectious clone was determined, and found to differ from our lab strain by only two silent mutations, one of which had been put in intentionally to create a restriction site. Thus, we presume the reduced specific infectivity of transcripts vs virion RNA is caused by errors during transcription of the 11 kb template and/or inefficient capping. Introduction of mutations in NS2B into the infectious clone. A series of mutants were made in which increasing portions of DEN2 NS2B was replaced by DEN4 NS2B. Only the mutant with the least amount of DEN4 NS2B grew at all, and it grew slowly and made small plaques. We plan to further characterize this mutant with regard to its growth properties in cells and in animal models.